A NEW BALANCE
between "Experience" and "Science"
Can feed Data = Examples = Experience into Example- ie Experience-Based Models, using techniques available from
Can setup High-Dimensional, Nonlinear Models of arbitrary complexity,
including "local explanation features"
to meet restricted low-dimensional human understanding capabilities.
Can do Assumption-Free Modeling with
Can integrate any Scientific, Rule-based,
and Heuristic Know-How
in one Example-Based Model.
Can develope from Discipline-Specific
to Interdisciplinary to Holistic Modeling,
using "mixed" (numeric and non-numeric) Parameter Sets.
Makes powerful Computer Aid useful
not only for Science-Based, but also for Experience-Based Modeling.
Can be used for High-Dimensional Nonlinear Parameter Identification,
to find most relevant Parameter Sets
for any given Problem.
Can be used for any application pattern,
Time Series Prediction,
Data Driven Processes,
Data Driven Businesses,
Data Driven Anything,
A NEW BALANCE
between "Experience" and "Science".
> Computer-Aided Example-Based Modeling (= CAEBM) can analyze and explain arbitrary, interdisciplinary, high-dimendinal problems from any domain, employing models with as well numeric as nonnumeric parameters.
> The associative modeling is done based exclusively on examples, using "Genetic Neural Nets". No theories, no assumptions, no prejudices, no additional verifications are required: The problem-examples alone, collected from any source, generate the models, with built-in analysis and representation of model complexity and accuracy.
> The resulting highly portable models can "explain" any relationship details between the participating parameters in 1 to 3 of many dimensions, making the models understandable by and plausible for "1-3 dimensional thinking" human users.
> As practically anything on earth can be seen as (a sequence of) examples, CAEBM can use the myriads of examples available and happening day by day, to detect the know-how contained, and to make it re-usable for humans and computers.
> In general, CAEBM can expand considerably the Experience-based problem understanding capabilities of humans, restricted normally to max 1-3 dimensions, the same way like Computer Aid (CA) has done it so impressively for Science-based problem understanding since several decades. A new fruitful balance between "Experience" and "Science" becomes possible.
> Know-how is the most important resource in today's businesses and organizations.
> Know-how is distributed in people's heads, in science-based models, in test results, in rules, in example-collections etc.
> Any know-how from any source, and anything (happening) in this world can be understood and represented by (a sequence of) examples, with numeric and nonnumeric parameters, if appropriate.
> Our Example-Based Modeling (EBM) technology constructs high-dimensional parametric models, from examples only. Neural Nets are used as modeling kernel and Genetic Algorithms setup the models and train them.
> We use Computer Aid (CA) for example preparation, iterative model setup, generalization & quality assurance, and parameter set optimization, resulting in our CAEBM technology: EBM+CA=CAEBM.
> Our results are handy computer models, containing the example-based know-how from any source and complexity in arbitrary problem domains, defined by the parameter set in use, error measures included.
> Our CAEBM technology can be used to collect, consolidate, continuously refine, and systematically RE-USE any know-how in any problem domain.
> Our CAEBM technology can construct new solutions for high-dimendional, interdisciplinary, even scientifically unresolvable problems, to make them understandable at the same time for human brains, normally restricted to max 2-3 dim problem understanding.
> Our CAEBM technology makes know-how becoming a portable, tradable product, independent from the know-how sources. This allows for totally new business opportunities, eg concentrating on know-how collection and refinement in their special problem domains. And until now "impossible" problem solutions become feasible.***
*** Design of predictive local, regional, national etc Covid-19 strategies, based on worldwide examples of infections and contra-measures taken, minimizing the damage done to people's health, to personal freedom, and to economy, while avoiding "wavy" pandemic developments.
*** Design of a low-cost, self-learning, local, regional, national etc weather prediction network, based on the weather itself as example source, making short to long-time predictions as needed at low budgets.
*** Design of a self-learning, inter-modal, local, regional, national etc traffic control network, based on the traffic itself as example source, minimizing traffic burden + maximizing transport performance for a given infrastructure, and identifying cost-minimal bottle-neck eliminations + most cost-efficient infrastructure improvements.
*** Setup of "intelligent", self-learning test stands, which collect their experiences gained so far by CAEBM technology in test stand-specific models (TSSMs), ready then to do additional test jobs then in a fraction of time, because most often new test jobs can be mostly fullfiled by the TSSMs, and only a few additional test stand runs are needed, to "calibrate" the experience to the new test job.
*** Setup and maintenance of a know-how network of CAEBMs for the development of a family of products (eg a family of cars), to be used for super-fast development (and production) of customer-individual products.
*** and many many more...